HPE Supercomputer Processors

HPE Processors for Supercomputers

Key Processor Series:

AMD EPYC™ Processors: HPE integrates AMD’s high-performance EPYC processors into their supercomputing solutions. AMD EPYC delivers up to 64 cores per processor, enabling ultra-high throughput and efficient parallel processing. With support for DDR4 memory and PCIe Gen 4, these processors are ideal for scientific simulations, AI workloads, and large-scale data analytics. Recent benchmarks have shown that the AMD EPYC 7003 series can deliver up to 3.5x the performance improvement in AI applications compared to previous generations.
Intel Xeon® Scalable Processors: HPE’s supercomputers powered by Intel Xeon processors offer unrivaled scalability with up to 40 cores per processor and advanced memory support. Intel’s deep integration of AI accelerators makes the Xeon series well-suited for research, automation, and deep learning models, which are key in industries such as healthcare and finance. Benchmark tests, such as those on the Intel Xeon Platinum 9200 series, indicate a 50% increase in throughput for deep learning workloads over previous Xeon generations.
HPE Custom HPC Processors: HPE’s own custom solutions are designed for specific industries, offering optimized performance for niche workloads. These processors are developed to meet the growing demand for specialized computing, such as real-time simulations in aerospace, automotive design, and financial modeling. These custom processors often integrate multiple accelerators to increase performance by up to 80% in targeted applications.
HPE Cray Shasta Architecture: The Cray EX supercomputers, powered by HPE Cray Shasta, leverage HPE's custom processors and accelerators for extreme performance and scalability. This architecture integrates processors with high-bandwidth memory and ultra-fast interconnects, enabling world-class performance for supercomputing applications. Benchmarking results from the Cray EX system, such as the installation at Argonne National Laboratory, demonstrate a peak performance of over 2 exaflops, making it one of the fastest supercomputers in the world.

Supercomputers Powered by HPE:

Perlmutter (USA): Named after Nobel laureate Saul Perlmutter, this system is built for AI and high-performance computing (HPC) workloads. Perlmutter leverages a combination of AMD EPYC processors and NVIDIA A100 GPUs to deliver exceptional performance in scientific research and simulations. It’s used at the National Energy Research Scientific Computing Center (NERSC) to assist with critical research in physics, climate modeling, and biology. Benchmarking Perlmutter's performance in climate modeling tasks has shown a 40% improvement in processing speed for simulations over previous generations.
LUMI (Europe): Located in Finland, LUMI is one of Europe’s most powerful supercomputers, offering over 550 petaflops of computational power. It is optimized for artificial intelligence, material science, and health data analytics, utilizing HPE’s advanced Cray EX systems integrated with AMD and Intel processors. LUMI’s integration of both AMD and Intel processors, combined with cutting-edge AI accelerators, allows it to deliver state-of-the-art performance for data-intensive applications.
HPC5 (Italy): HPC5 is Italy’s most powerful supercomputer, with a unique focus on energy and environmental sustainability. Powered by HPE Cray and AMD EPYC processors, HPC5 is designed to simulate complex energy and environmental systems, supporting industries such as oil exploration and clean energy research. HPC5 has been benchmarked to provide optimal simulation capabilities for energy grids, reducing simulation time by over 50%.
Red Sky (USA): Located at NASA’s Ames Research Center, Red Sky uses Intel Xeon processors to provide the processing power needed for critical aerospace research, including simulations for spacecraft designs, satellite operations, and deep space exploration. With its optimized architecture, Red Sky has been instrumental in reducing the computational cost of spacecraft trajectory simulations, cutting runtime by 30%.

Key Features of HPE Supercomputer Processors

Exceptional Scalability: HPE processors support scalable performance for workloads ranging from small research projects to large global simulations. HPE’s flexible architecture allows researchers to start with smaller systems and scale up as their computational needs grow. Benchmarks show that scalability on the Cray EX architecture increases by over 60% in multi-node performance.
Advanced AI and ML Optimization: HPE supercomputers come equipped with processors optimized for AI and machine learning (ML) workloads. The combination of high core counts and AI acceleration technologies ensures that training deep learning models and performing data-intensive tasks can be done efficiently and at scale. For example, the EPYC 7003 series processors show up to 2x better performance in AI workloads compared to previous models in tests like MLPerf.
Low Latency and High Bandwidth: HPE processors feature fast memory and high-bandwidth interconnects, providing low-latency communications between nodes and GPUs. This ensures minimal bottlenecks and maximizes performance for complex workloads. Benchmarking on Cray EX systems demonstrates a latency reduction of 25% when running interconnect-heavy applications.
Energy Efficiency: Built with energy efficiency in mind, HPE processors deliver powerful performance while maintaining low power consumption. HPE’s energy-efficient architecture helps reduce operational costs and minimizes the carbon footprint of supercomputing facilities. Benchmarks indicate that the Cray EX system delivers 20% greater performance per watt than previous supercomputer generations.
Security and Reliability: HPE processors are designed with built-in hardware security features, including secure boot, memory encryption, and advanced error-correction technologies. This ensures the integrity of sensitive data and mission-critical applications running on HPE systems. The Xeon Scalable processors, for instance, include support for Intel's Total Memory Encryption (TME) and secure launch technologies.

Applications of HPE Supercomputer Processors

HPE supercomputers are transforming industries by providing unparalleled computing power for a variety of critical applications:

Climate Change Modeling: With unprecedented computational power, HPE supercomputers are used to run complex climate simulations, forecasting environmental changes, and analyzing the impact of global warming on ecosystems and human populations. The Perlmutter system's role in climate modeling has helped accelerate predictions on the effects of climate change by a factor of three in some scenarios.
AI and Data Analytics: HPE supercomputers accelerate AI/ML workloads, enabling industries such as healthcare and finance to analyze massive datasets in real time. This includes predictive modeling for disease outbreaks, fraud detection, and optimizing supply chains. LUMI, for example, has been benchmarked to cut training time for AI models in half, significantly speeding up machine learning research.
Genomic Research: Supercomputing capabilities are being used to process and analyze genomic data, assisting researchers in drug discovery, genetic research, and personalized medicine. HPC5’s capabilities are pushing the boundaries of genomic analysis, enabling faster drug discovery timelines by as much as 40%.
Automotive Design and Simulation: Automotive manufacturers use HPE-powered supercomputers to simulate vehicle designs and optimize manufacturing processes, improving fuel efficiency, safety, and performance. Benchmarking of automotive simulations on Cray EX systems has shown up to a 50% improvement in simulation speeds.
Aerospace and Defense: Simulations run on HPE supercomputers assist with aerospace design, satellite tracking, and defense applications, enhancing national security and space exploration missions. Red Sky’s contributions to aerospace modeling have shortened satellite design cycles, reducing time-to-market for critical aerospace technologies.

HPE Processors for Supercomputers

Key Processor Series Benchmarks

Processor Series	Key Features	Benchmark Metrics	Performance Improvement
AMD EPYC™ 7003 Series	64 cores per processor, DDR4 memory, PCIe Gen 4	Up to 3.5x better performance in AI workloads	Performance improvement in AI tasks compared to previous generations
Intel Xeon® Platinum 9200 Series	Up to 40 cores, AI accelerator integration, advanced memory support	50% improvement in deep learning throughput	Significant boost in AI/ML workload performance
HPE Custom HPC Processors	Custom-designed processors, tailored for niche industries	80% improvement in specialized workload performance	Custom solutions outperform general processors in targeted applications
HPE Cray Shasta Architecture	High-bandwidth memory, ultra-fast interconnects	Peak performance: 2 exaflops	Extreme scalability and performance for world-class simulations

Supercomputers Powered by HPE

Supercomputer	Location	Processor Used	Benchmark Performance
Perlmutter	USA (NERSC)	AMD EPYC™ & NVIDIA A100 GPUs	40% improvement in climate modeling simulation speed
LUMI	Finland	HPE Cray EX with AMD & Intel processors	Over 550 petaflops of computational power, optimized for AI
HPC5	Italy	HPE Cray & AMD EPYC processors	50% reduction in simulation time for energy and environmental systems
Red Sky	USA (NASA Ames Research Center)	Intel Xeon processors	30% reduction in spacecraft simulation runtime

The Future of HPE in Supercomputing

HPE’s vision for the future of supercomputing includes continued innovation in processing power, AI integration, and green computing. The future of HPE processors will focus on enhancing AI and ML capabilities, pushing the boundaries of quantum computing, and offering even more energy-efficient systems that contribute to sustainable supercomputing.

In the next decade, HPE aims to create processors with even higher core counts, greater memory bandwidth, and the ability to seamlessly integrate into exascale and quantum computing systems. HPE’s investment in AI and quantum technologies will allow the company to support a new era of supercomputing, solving some of the world’s most pressing problems, from curing diseases to advancing space exploration.

HPE PC Processors

About HPE in Supercomputing

HPE Processors for Supercomputers

Key Processor Series:

Supercomputers Powered by HPE:

Key Features of HPE Supercomputer Processors

Applications of HPE Supercomputer Processors

HPE Processors for Supercomputers

Key Processor Series Benchmarks

Supercomputers Powered by HPE

The Future of HPE in Supercomputing